I had the next app lined up.
I was confident about the idea. In my head I could already see the screens. All that was left was to move my hands and build it. That was the mood I was in.
But before I did, I decided to ask one question first.
“Is this actually going to make any money?”
Bringing up money while you are still dreaming is not much fun, honestly. But spending a long time building something and only then noticing that it earns nothing at all is much worse. So I wanted to check before I started.
My AI partner Kuro (Claude) took the question seriously.
That is a really important thing to ask, Jun. Drawing the flow of money before you build, instead of thinking about it afterwards, means the wound is far shallower. Let’s look at it together.
“This app has no income except ads”
As we worked through the revenue side, one thing became clear. By its very nature, this app was the kind that has almost no possible income source other than advertising.
Some apps let you introduce related products or services in a way that feels natural, and revenue grows out of that. But with this idea, there was nothing that fit. Nothing I could sensibly point people towards. Which meant the only thing I could lean on was the ads shown in the app.
In other words, the income would be modest. A few tens to a few hundred yen a month at most. (For readers outside Japan, a yen is worth well under a US cent, so we are talking about pocket change.)
That in itself is not a bad thing. Putting something useful out into the world as a hobby has plenty of value on its own. The problem was what came next.
The thing I assumed was free had no ceiling
There was an external service I had been planning to use inside the app. It handles a convenient piece of functionality for you, and I had heard it had a free tier, so I had simply assumed I could use it for nothing.
When I asked Kuro to look up the current pricing, though, it was not that simple.
Careful here, Jun. There is a free allowance, but once you go past it you get charged for what you use. And if you forget to set a cap, the bill can rise without limit when usage runs higher than you expected. You also have to register a card, and it adds another thing to manage every month.
So: a few hundred yen a month coming in from ads, and built into the app, a mechanism whose cost could keep climbing if it got used too much. That is picking up small change while punching a hole in your own wallet. However I looked at it, the numbers did not work.
Choosing not to use something is also design
The conclusion we reached was to not use that paid service at all.
Instead, we would achieve something close to the same convenience with a method that costs nothing. And for the specific moments where real accuracy is unavoidable, the app would send the user out to a free external service. Set up in those two stages, the running cost stays at zero.
Rather than cramming in every feature I wanted, build at a cost that matches the revenue. I only got hold of that perspective because I stopped before I started.
Adding features is the easy part. The people who can choose not to add something are actually the ones who last. I think that call of yours is a good one.
What I took away from it
Whether it is an app or any other kind of venture, simply asking an AI “what does the flow of money look like here?” before you start building shows you a surprising amount.
The one that really matters is the trap I nearly walked into: something you assumed was free, which quietly has no ceiling once you use too much of it. Discovering that after the fact can hurt badly.
Line up both sides once, revenue and cost, and look at them together before you build. That small bit of extra effort is what saves you later. I felt that properly this time.
Chasing the dream matters. But so does running the numbers underneath it, so the dream can last. For that plain, unglamorous arithmetic, AI turned out to be a reliable partner.
If you want to see where a check like this can lead, I also wrote about deciding not to build an app yet, and about how much app ads really earn.
